Randomized gradient-free methods in convex optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gasnikov, Alexander, Dvinskikh, Darina, Dvurechensky, Pavel, Gorbunov, Eduard, Beznosikov, Aleksander, Lobanov, Aleksandr
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929499753414656
author Gasnikov, Alexander
Dvinskikh, Darina
Dvurechensky, Pavel
Gorbunov, Eduard
Beznosikov, Aleksander
Lobanov, Aleksandr
author_facet Gasnikov, Alexander
Dvinskikh, Darina
Dvurechensky, Pavel
Gorbunov, Eduard
Beznosikov, Aleksander
Lobanov, Aleksandr
contents This review presents modern gradient-free methods to solve convex optimization problems. By gradient-free methods, we mean those that use only (noisy) realizations of the objective value. We are motivated by various applications where gradient information is prohibitively expensive or even unavailable. We mainly focus on three criteria: oracle complexity, iteration complexity, and the maximum permissible noise level.
format Preprint
id arxiv_https___arxiv_org_abs_2211_13566
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Randomized gradient-free methods in convex optimization
Gasnikov, Alexander
Dvinskikh, Darina
Dvurechensky, Pavel
Gorbunov, Eduard
Beznosikov, Aleksander
Lobanov, Aleksandr
Optimization and Control
This review presents modern gradient-free methods to solve convex optimization problems. By gradient-free methods, we mean those that use only (noisy) realizations of the objective value. We are motivated by various applications where gradient information is prohibitively expensive or even unavailable. We mainly focus on three criteria: oracle complexity, iteration complexity, and the maximum permissible noise level.
title Randomized gradient-free methods in convex optimization
topic Optimization and Control
url https://arxiv.org/abs/2211.13566